If you haven’t noticed already, there was no simple way to just enter a web URL in SCAN and get a document in the repository. One had to add web documents via del.icio.us or fetch them from RSS. Now it is fixed – with Web plugin you can add web-pages with their URL’s and more – you can tell the plugin to follow hyperlinks and crawl web-pages recursively, thus adding whole web-sites to SCAN document repository.

More on Web-plugin »

To install new plugin, click on “Check for updates” button in Plugins manager window (“Tools?Manage plugins”). Web plugin will appear in “Location type plugins” section.

A day in the blog’s life.

It’s not a 04-01-joke, it’s like the Slashdot effect looks. Someone mentioned my old article (“Anicent tags museum“) in a comment to the Slashdot post and it resulted in a traffic that Chronicles usually have during a whole month.

I am not an active user of the core Y! services, but I’ve got a Flickr and Del.icio.us. Though they are important parts of my everyday online life, I definitely will have to go off if MS would be there. At least, that’s good that I’ve not purchased Flickr Pro account yet, so they will not get my hard-earned cents!

Who knows a good photohosting service with possibility of import all shots from Flickr?

This decision seems to be related with recent turn of portal orientation towards a community-oriented service. As SourceKibitzer’s CEO Mark Koffman told me, “we are moving from software towards the people behind it”. Recently, the portal has launched new features to support this direction, namely online resume for Open Source community members (Bio) and series of interviews with OSS developers.

I think, it has good chances to grow into a real social network for the FOSS folk.

At the moment, the portal code is published under GPL v3 Draft. An exception is the proprietary libraries implementing the benchmarking algos, which are know-how of the company running SourceKibitzer. The portal invites everyone to join the development by subscribing the mailing list , participation in TRAC Wiki and contributing the code into SVN repository. The portal technology is based on Java Spring framework.

I expected that Firefox would can update itself via built-in updating mechanism, but it was not so. The latest version which the updater
could find was 1.5.0.7. So, it was needed to download and unpack a tar.gz distributive from GetFirefox. I thought it was even better way to install, because the previous working installation was left untouched and I could roll back safely if something were go wrong with 2.0 version.

Luckily, nothing bad was happened. Installation (which is just the first Firefox launching) has gone smoothly and without any problem. Firefox transparently imported all my settings, bookmarks, toolbars and history, so nothing had been lost from my previous installation. It also checked the installed extensions for compatibility with 2.0 version and offered to install newest versions of outdated (incompatible) add-ons. The most important ones (such as del.icio.us, Adblock, FlashGot etc) already had 2.0-compatible updates so there were no problems with them. In fact, there was only one add-on which caused the problems – a custom extension by my ISP for displaying my account balance in Firefox statusbar. It was very helpful and I didn’t want to lose it even until an updated release. Fortunately, two-minutes-long research and a little hack made it 2.0-compatible :-) – I just changed maxVersion parameter in its install.rdf manifest file.

So, besides the smooth and easy installation, there is a lot of improvements I liked:

• Session support is a great new feature – you can close the browser with a heap of tabs opened, turn your machine off and go to bed. The next time you started Firefox, you will get the same set of tabs with the same web-sites and start to continue your work immediately.
• New tabs are more usable – now they have the closing buttons and site icons. When it is too much of the tabs opened, they can be scrolled. It seems that the developers have been inspired by Eclipse tabs – there is exactly the same user experience.
• Built-in spellchecker is awesome. Actually, it already helps me when I’m typing this article.
• New add-ons manager now combines both extensions and themes management and updating.
• It is possible to set a custom application to handle RSS links. Unfortunately, it doesn’t work for me, as Akregator (my RSS client) needs “-a” option to add a new feed, but Firefox doesn’t allow to edit a command line. But it should work fine with on-line agregators (and few ones are predefined there).

Well, Firefox 2.0 is, of course, not a revolutionary step in web-browsing experience. It is just an usual best browser on the planet which now is got even better. It’s just Firefox.

[added, 10-30]

My software reviews usually have two parts “What’s good” and “What’s bad”. As this review had been written under first impressions, I had no idea what’s bad in Firefox 2.0. But after three days of using it I found a really nasty bug: When scrolling the window, there are the visual glitches with fonts. Sometimes they look just mangled. I have no idea is it an X, xfs, or Firefox issue, but I didn’t see it on 1.5 version. So, “What’s bad” section is opened. Sadly enough.

There is a case study: our RUSMECO project, among many other things, includes development of so-called “Brokerage Area” – an open-access database of the people, organizations, products and services which are related together by many ways. Our main collaboration portal is based on Drupal so it was obvious that reusing the same platform for BA development advantages in all senses. We have investigating which Drupal features might fit our solution – and found all what we needed.

Flexinode

All content objects in Drupal are treated as an abstraction of the “nodes”. There are various types of nodes (pages, stories, forum posts, blog posts etc) which differ by their fields and provided by the core or additional modules. But there is also a mechanism of creating new custom types of nodes (so-called “flexinode”) with arbitrary sets of the data fields. So, you can define your own structured data types right into the Drupal administrative interface without writing any PHP code and creating new database tables. The user interface (the forms) for nodes will be built automatically and can be adjusted later (for instance, the fields order and grouping). These new types will be fully integrated into the core content framework – in fact, they will be the same content objects as the core types and all Drupal functions (the search, categorization, multilingual translation, workflow, access management and so on) will work with them.

In our case, we needed four content types: Member, Company, Product and Service profiles. It is only pity that the built-in system user profile is not a node (though it have a customizable set of the fields), so we left there only a bare minimum of fields (only username, email and password, in fact) and defined a separate Member profile (CV) node type for those who wants to tell more about herself.

Taxonomies

Taxonomies (or, controlled vocabularies) is a standard Drupal way for nodes classification and organization. They are the sets of categories or terms which can be organized hierarchically, related each other and grouped into the vocabularies. Depending on your IA, they can serve as traditional content sections, the “tags” or “labels” in faceted classification system or as thesauri of interrelated terms. The categories can be also related with the content types, so that the nodes of a specific type can be qualified as the members by specific categories only, or be included in a category by default.

Along with predefined categories, it is possible to define new ones (in selected vocabularies) on-the-fly, together with creation of new nodes. For instance, an user who is filling her product profile out, is able to create the new product category at one time.

I18N

Our solution had to be bilingual (en/ru) – it was a principal requirement. It concerns not only an interface, but also that it should be possible to create the translations of the nodes in alternate language. For instance, an user should be able to create two versions of a profile and when another user coming to see it, she would get a variant for her current language (selected in the top panel).

The features of Drupal i18n module enable to do so. What’s more, it is possible also for categories to be translated, so that the user sees all categories in her current language and when she is in a category, all nodes there are in that language too.

Modules

And after all, if you’re well sure Drupal lacks some functionality, you can remember your own PHP and SQL experience and write some code. For instance, you would be needed for a custom sophisticated query iinterface for your database. But you probably will not have to write a lot – as all infrastructural aspects (such as templating etc) are already implemented in Drupal API, you’ll need to implement only your specific functions.

The specific pieces of Drupal functionality are called “the modules”. In fact, the Drupal core is only the component framework and almost all its functions are implemented as the pluggable modules. So, to extend it with your custom functions there is no a better way than develop your own module. It’s pretty easy – a module is simply a PHP file containing the conventionally named functions (aka “hooks”). These functions may define what should be displayed in a module page, module blocks (in left and right columns), module settings page, how the module will be integrated into the navigation menus and configuration interface and so on.

For to start with modules development, consider to read this tutorial.

Reading on subject

• Is Drupal right for you?
• Rolling your own system vs. using Drupal
• Drupal for Information Architects: an overview of configurability
• The Road to Drupal Hell

… more on the Drupal site.